最新刊期
卷 11 ， 期 1 ， 2006
- 摘要："Journal of Image and Graphics" started its publication in 1996. For 10 years,it has made important contributions to the progress of the researches applications and developments of image and graphics in China.This paper according to the principles and methods of bibliometrics made a systematic investigation and analysis on the articles published in this journal and the authors of these articles in the last 10 years. This work shows the progress trace of "Journal of Image and Graphics" after its initiation and reveals the current status of researchers on image and graphics.It not only affirms the rationality of this journal as a core journal in the science of image and graphics but also provides useful reference data for the further development of the science and advancement of the journal.3156|205|0更新时间：2024-05-07
- 摘要：Segmentation of merged characters is one of difficulties that have attracted a great deal of attention in optical character recognition(OCR).Nowadays,unsuitable segmentation is the primary cause of recognition errors.Based on the analysis of the shortcomings of some traditional algorithms for printed character segmentation,we notice that it is necessary to propose a fine method to meet with the requirement of real time processing and the characteristics of DSP module including relative low power and small memory comparing with PC.In this paper a new algorithm of segmentation and recognition based on connected component is proposed which can be used for electronic reading-pen.The proposed method reduces computation time by recognizing the connected component as a whole.It segments connected component by middle expansion method and peak-paddle function.As a result,recognition error arose by segmentation error can be reduced.Experiment results have proved that the algorithm is effective,easy to implement and it is reasonable and applicable for to ERpen.关键词：character segmentation;connected component recognition;electronic reading-pen3228|224|0更新时间：2024-05-07
- 摘要：In order to enable traditional monitor systems to detect and track moving objects, a statistical model and GVF- Snake based algorithm is proposed, which employs color video information in static background to detect and track moving objects and provide a good representation of the objects so as to simplify the subsequent object recognition. The algorithm proposed replaces the plain gray space model with normalized RGB space combined with gray space model to eliminate the effect of shadow upon detection, A GMM of the difference of 2 successive frames is constructed, based on which, a motion border image is generated. GVF-Snake is then enhanced to extract the contours of moving objects in video sequence by modifying the energy entry and adding a method to initialize the Snake automatically. In order to accelerate the convergence of the Snake, the contour of next moment is predicted at every moment by estimating the center of the moving object using a 1st-order difference algorithm. This algorithm has been proved to be effective for both rigid and non-rigid objects and can be used for smart surveillance and traffic monitoring,2843|278|0更新时间：2024-05-07
- 摘要：Face verification is essentially a problem of one-class classification or outlier detection.Its goal is to accurately describe one class of objects opposing to a wide range of other objects considered as outliers.Based on the existing research wook on statistical learning theory and kernel methods,this paper analyses the drawbacks of a existing one-class classification algorithm,namely support vector data description(SVDD),on dealing with dynamic samples of face verification.This paper points out that,with the increase of samples,the size of optimization will exceed the memory space of the computer.Consequently,the algorithm will be unable to carry out.For the purpose of reducing the size of optimization,the dynamic support vector data description algorithm(DSVDD) is proposed.The new algorithm only computes the new samples and support vectors in the process of optimization,so that the required operation size and memory space can be reduced in a great degree,which means the real-time and dynamic demands are met for face verification.关键词：data description;face verification;kernel methods;outlier2511|180|0更新时间：2024-05-07
- 摘要：Accurately locating the human eyes in a single and gray level image with cluttered background remain a difficult problem in recent year.In this paper,we use a discrete Hidden Markov Model(HMM) that included explicit state duration density to locate the human eye precisely.At first,a retinal sampling grid and a special method are adopted to extract the observation sequence.We do not need rotation and scaling transformation and matching operation,for we use a dynamic observation symbol probability distribution in state to adapt the various angles of human eyes in image,and a dynamic sampling model that controls the sizes of the retinal sampling grid through the evaluation of the observation sequence to adapt to the various sizes of human eyes in image.The experiment results show that our algorithm is effective and robust and have high positioning accuracy.关键词：hidden markov model(HMM);localization of human eyes;face detection3053|240|0更新时间：2024-05-07
- 摘要：Local Fourier transform(LFT) is an important texture feature extracting method.In this paper,the texture discrimination performance of LFT coefficients' moments is studied.The Fisher criterion is used to evaluate the discrimination performance of different texture features.Texture image segmentation experiments enable the conclusions more reliable.It is revealed that the LFT coefficients of texture images usually do not obey normal distribution,their even-order moments exhibit preferable texture discrimination performance,while their odd-order moments exhibit baddish.If LFT coefficients' 2nd,4th and 6th order moments are used as texture features,better texture discrimination performance and segmentation results can be achieved than those of Yu Hui and Haralick.So,it is suggested that the even-order moments of LFT coefficients are selected as texture features.关键词：image processing;pattern recognition;texture features;texture analysis;local Fourier transform2693|255|0更新时间：2024-05-07
- 摘要：Basic conception and correlative properties of multiple order morphology transformation on grayscale image were introduced.The principle of edge detection using order morphological filtering was illustrated.The effect of structuring elements and percentile on edge detection was discussed.Based on the idea of multi-scale morophological filtering,three order morphological operations of edge detection were constructed by extending the basic morphological operations of edge detection to restrain noise.The specialities of the operations which were monotony about percentile(p,q) in the area of image edge were analysed in theory.Based on above,three general order morphological edge operations were constructed and their formats were given,in which multi-structuring elements were selected to match image edge according to geometrical feature of image and to remove noise in diffenent directions in the meanwhile.The rules of how to select structuring elements and percentile(p,q) of duplicate mixed order morphology transformation were studied.The experiments showed that the operations we presented were more effective than other classical and ordinary morphological operations of edge detection on noise restraining and retaining the image details.关键词：order morphology filter;structuring elements;edge detection;noise restraining3078|214|0更新时间：2024-05-07
- 摘要：Skin color has been proven to be a useful and robust cue for face detection,localization and tracking.Naked skin region is one of the most important features to detect the erotic pictures.For detecting the skin images effectively,a skin color classification technique that employs Bayesian decision with color statistics data,which include(1 809 502) skin pixels and(1 763 682) non-skin pixels gathered from YCbCr color space,has been presented.Considering the influence of luminance(Y component) on skin color,the skin color model is constructed by two lookup tables in Y-Cb and Y-Cr subspaces.To use 2 lookup tables(LUTs) together,Gaussian normalization and linear converting are used to normalize the range of threshold value to.To filter scattering points,both LUTs are processed by median filter.Experimental results have demonstrated that this method can provide lower false dismissal rate(9.814%) lower false detection rate(3.5%),compared with three methods given in this paper,and achieve better detection performance in various lighting conditions.关键词：Bayes decision;lookup table(LUT);Gaussian normalization;skin color detection3398|278|0更新时间：2024-05-07
- 摘要：The registration of sequential images acquired through floating aerial platform is a difficult problem up to date. The floating aerial platform can get sequential images of the down-looking scene in bad environment, with which the wide visual field image can be got by stitching for further research . Influenced by many factors, sequential images have the transformation of rotation, unequal scaling and translation, noise and geometric deforming. In order to solve this problem, this paper proposes a new algorithm. In the algorithm, firstly, segment the image with the iterative threshold method and use crossing linker to record and express the characteristic area information. Secondly, construct the homonymy matrix of characteristic areas and search the homonymy matrix to get matching pairs of homonymy areas between sequential images automatically, Lastly, calculate the registration parameters of sequential images including rotating, scaling and translation. The experimental result demonstrates this algorithm is robust, efficient and reliable for certain amount of sequential images, and has theoretic value and practical significance.关键词：image stitching;sequential frame image;homonymy degree;homonymy matrix2712|195|0更新时间：2024-05-07
- 摘要：Image enhancement is an important task of image processing. We propose an algorithmic adaptive image enhancement method based on image information measure in this paper. At the image pixel level, image pixels are classified into smooth pixels and edge pixels by using image information measure, then the smooth pixels are farther classified into smooth pixels and ridge pixels aiming at the feature of form image. Relevant filtering algorithms are designed for different types of pixels, which realizes algorithmic adaption and enhance image primarily. Finally, the multi_level fuzzy enhancement algorithm is imposed to enhance image contrast and achieve a better enhancement effect. Experimental results show that the algorithm can enhance the low contrast image effectively and obtain a very good visual effect.关键词：image enhancement;image information measure;step edge;ridge edge;pixels classifying2668|210|0更新时间：2024-05-07
- 摘要：Although there are varieties of image segmentation algorithms, no one is applicable to all images. In order that the image tracking system can select segmentation algorithm itself according to the object image feature, this paper presents an intelligent decision method Of image segmentation. Firstly, some representative segmentation algorithms are selected to form an algorithm library, using which various sample images are segmented; Secondly, decision information table is builtup based on diversified numerical features extracted from the sample images and the optimal segmentation algorithm of each sample image judged according to segmentation quality evaluation criterion; Finally, rough set theory is applied to diseretization and attribution reduction of decision information table, in order to make the decision rule of image segmentation algorithm seiection. The decision method solves a series of problems for segmentation algorithm selection in image tracking system. As experiment shows, it can effectively pick out the optimal segmentation algorithm from algorithm library according to the feature of the processed image, and also satisfy the real-time demand of image tracking system on vehicle.2597|218|0更新时间：2024-05-07
- 摘要：This paper dedicates to develop an algorithm which can be used to detect the real-time distance between the own car and the obstacle ahead. On the basis of comparing several existing methods, this paper uses the geometrical reasoning method to obtain the depth information of the driving environment for the vehicle control, and presents a real-tlme monocular-vision-based algorithm to get the distance between the own car and the obstacle ahead. The pitch angle of the camera is the key factor which affects the precision of the distance detection. This conclusion is drawn from the experimental results. In order to get the real pitch angle of the camera, the constraint that the two lane boundaries are parallel is used in the algorithm. The results of the static real-car tests show that the algorithm presented in this paper satisfies the precision requirements of the distance detection. And the results of the dynamic real-car tests show that the algorithm also satisfies the real-time requirements for the intelligent vehicle control.关键词：monocular vision;depth information;boundary restriction;real-time distance detection4258|284|0更新时间：2024-05-07
- 摘要：The colonoscopy heavily depends on the proficiency of the operator at present.For achieving an accurate and safe colonoscopy,the intelligent robotics is applied here for visual navigation of the colonoscopy.In this paper, a new visual navigation method for colonoscopy is described and tested,which is based on analyzing colon contours.For acquiring the precise edge energy of the colon contours,it uses steerable filters to detect the edge energy and the directions of the colon contours.Then it accumulates the edges' energy mapped in the normal directions of the edges.And by searching the area with highest accumulated energy,this method finds the center area formed by the curved contours. This method is designed to assist the visual navigation based on dark area analysis for reducing the possibility of navigation failure,which may be caused by the unsteadiness of the dark area.Some experimental results have demonstrated the effectiveness of this method for the assistance to the visual navigation for colonoscopy.关键词：colonoscope navigation;colon contour;steerable filter2854|204|0更新时间：2024-05-07
- 摘要：A curvature adaptive algorithm is presented to simplify 3d meshes rapidly based on splitting of the parameterization plane. The original 3d mesh is mapped onto the parameterization plane. A planar gray-level distribution is constructed based on the curvature values of the vertices of the original mesh to proclaim the various sampling density requirements of the original mesh. An iteration algorithm is used to select the curvature adaptive resample vertices on the parametcrization plane to construct the simplified mesh. The parameterization plane is iteratively split into two parts with the same summation of the gray-level value in each step. A non-balance binary tree was constructed during the procedure of splitting. The resample vertices set is obtained by allocating one vertex in each leaf node, and those resampled vertices are triangulated to construct the simplified mesh. This algorithm is very fast and can preserve the detail very well.关键词：mesh simplification;parameterization;curvature2742|238|0更新时间：2024-05-07
- 摘要：The key point in smooth and fairing fitting of curves and surfaces is to search the optimal parameters of data points, and then construct an optimal fitting equation according to the least square method and get control points based on the equation. As existing techniques for choosing parameters do not embody the geometric characteristic of the optimal parameters, the fitting is either imprecise or of great cost. In order to improve fitting precision and computing speed, we offers an algorithm on optimizing the parameters of data points. By using the orthogonal projection of data points to corresponding curve or surface, and making a search for the neighborhood of the parametric coordinate computation is speed up and the parameters can be ceaselessly corrected with the iterative process of the curves and the surfaces, so that the resulting parameters will possess distinct geometrical meaning, and the optimal fitting effect can be obtained, Comparing with the algorithms of Hoschek, Carlos and Picgl in some examples, it is validated that this method can cut down iteration times about 10% - 90% , reduce time consumption around 20% -70% , or improve precision nearly 40%.关键词：curve;surface;fitting;parameter;orthogonal projection2666|195|0更新时间：2024-05-07
- 摘要：Contour line is a kind of map symbols, with which 3D terrain can be recorded on 2D map. Map users know the elevation of terrain with the change of contour line. The changes of contour line means local change and whole change. The latter is also called shape change, and it contains more information of terrain. There are many studies on local change of contour line, and a lot of achievements have been made, such as curvature analysis, dominant points detection, and so on. There are little studies on whole change of contour line, which leads to some problems related to whole terrain remaining unsolved, After studying on shape change of contour line deeply, the author concludes the similarity rule for shape change of contour lines. It provides a new way to interpolate contour line with two neighboring contour lines, Besides, the methods put forwards in this paper can be used to reconstruct 3D terrain surface with contour data. Experimental results show that the new methods can interpolate contour line effectively.关键词：similarity;contour line;interpolation model;3D terrain2799|191|0更新时间：2024-05-07
- 摘要：Been used special optic method to produce independence view zone for each eye, made the stereo pair corresponding with left and right eye, the nakedness-eye stereoscopic display can be perceived stereo vision without stereo spectacle. In this paper, we briefly several related typical stereoscopic display using TFT LCD. And taking the line light illumination 3 D display as main research object, we analyzed the section light road. By using exposal method to make grating film, we designed a linear light illumination board. Moreover we discussed basic structure of stereoscopic display. The result of experiment proves our stereoscopic dispIay can perceive high resolution, no twinkling stereo vision.关键词：thin film transistor liquid crystal display(TFTLCD);stereoscopic;linear light;film2810|229|0更新时间：2024-05-07
- 摘要：Comparing with the traditional approach of 3D model-based rendering with explicit 3D structure, the approach of image based modeling and rendering has many advantages, but both fall futile in modeling from one single picture. The method of touring into the picture, proposed by Horry in Siggraph 1997, gives us a surprising way to solve this problem. Our methodology in modeling from one picture can be recognized as an extension to TIP. Based on these extensions on TIP, we consider an image sequence taken with the unknown path of the viewpoint. The path can be a line, curve or a combination of both. Each frame of the sequence is decomposed into sub-images, which are mosaicked into a new texture to be used in the following walking through. Wherever the viewpoint is, it can get high-resolution virtual images. The experimental results show that this method is effective and applicable,关键词：tour into a picturel(TIP);image sequence;image mosaic;walk through2506|176|0更新时间：2024-05-07
- 摘要：As an important image-processing tool, Hilbert curve is widely used in image processing, especially in image scanning. To produce the curve correctly and quickly, a new algorithm of Hilbert curve based on matrixes is presented with its MATLAB program, and is proved by means of mathematic induction. The latter numeral simulations also illustrate the correctness and feasibility of the algorithm in the paper. An application example of the algorithm is given in the end.关键词：Hilbert curve;image scanning;MATLAB4439|263|0更新时间：2024-05-07
- 摘要：Huge amounts of data of hyperspectral images have become a great challenge to data storage and transmission.Due to the special way to obtain hyperspectral images,they are costly,but this kind of image can be used extensively in the meantime.Any information shouldn't be lost during compression,so an efficient lossless compression method seems to be essential.Without an efficient compression scheme,the application of hyperspectral images will be restricted.Trellis coded quantization(TCQ) inherits ideas from trellis coded modulation(TCM) the expanded signal set,set partitioning and trellis state transition.The mean squared error(MSE) performance of TCQ is excellent with modest computing complexity.At present TCQ is mainly used for lossy image compression.In this paper,we aim to develop TCQ scheme to compress hyperspectral images losslessly and find a new approach yielding higher lossless compression ratio.We take advantage of the characteristics of hyperspectral data for compression.A scheme utilizing wavelet transformation and TCQ is presented.Compared with the result of JPEG2000 and JPEG-LS,experiments show that the method presented in this paper has better performance.关键词：hyperspectral imagery;lossless compression;wavelet transformation;trellis coded quantization(TCQ)2766|199|0更新时间：2024-05-07
- 摘要：Long coding time is the main problem in fractal image compression at present,to which classification of image blocks is an important and efficient solution.However,there is an contradiction between coding speed and image quality in most classification methods.Aiming at the contradiction we first give out the guide lines of the performance for classification methods,hence propose a technique called "Adaptive Classification",which is then used in the classification by mass center combined with satisfied match,thus obtain a fast method for fractal coding.Experimental results indicate that at the same compression ratio,the proposed method improves the speed of fractal coding and the quality of decoded image,in contrast with the original uniform classification.关键词：fractal image compression;iterated function system;adaptive classification2665|186|0更新时间：2024-05-07
- 摘要：This paper presents a joint source-channel coding scheme based on SPECK and revises turbo codes according to the characteristic of source and channel coding.The purpose of this scheme is to improve the robustness of the image.Source coding in this paper adopts image coding based on wavelet.This scheme uses SPECK to generate diverse fault tolerant substream.Punctured Turbo codes are used for channel coding.Channel coding adopts unequal error protection(UEP) for the compressed data with different channel coding rates according to the importance for reconstruction.This scheme utilizes the property of compressed data which source has coded and gets better balance between bit error rate and code length.The experimental results show that the proposed scheme can get high PSNR(peak signal noise ratio) in high compression ratio and provide strong robustness to the image in low SNR because of the superior performance of Turbo-codes.关键词：image coding;set partitioned embedded block coder(SPECK);joint source channel coding;unequal error protection(UEP);Turbo codes2819|217|0更新时间：2024-05-07
- Postal code：100190
- Tel：010-58887035/58887030/58887418 Email：jig@aircas.ac.cn
- Technical support is provided by Beijing Founder electronics co., LTD 京ICP备05080539号-4
京公网安备11010802024621 - It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
- Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.
0